Hepatocellular Carcinoma in Children: Results of the First Prospective Study of the International Society of Pediatric Oncology Group

By P. Czauderna, G. Mackinlay, G. Perilongo, J. Brown, E. Shafford, D. Aronson, J. Pritchard, P. Chapchap, J. Keeling, J. Plaschkes, and J.B. Otte for the Liver Tumors Study Group of the International Society of Pediatric Oncology
Purpose: To improve survival and reduce operative morbidity and mortality in children with primary epithelial liver tumors by using preoperative chemotherapy, as well as to collect information on the epidemiology, natural history, and prognostic factors. Patients and Methods: Forty children with hepatocellular carcinoma (HCC) were registered onto the Group for Epithelial Liver Tumors International Society of Pediatric Oncologys rst study from January 1990 to February 1994. The outcome could be analyzed in 39 of those patients. Disease was often advanced at the time of diagnosis; metastases were identied in 31% of the children and extrahepatic tumor extension, vascular invasion, or both in 39%. Multifocal tumors were common (56%). Thirty-three percent of tumors were associated with hepatic cirrhosis. All but two patients received preoperative chemotherapy (cisplatin and doxorubicin). Results: Partial response was observed in 18 (49%) of 37 patients; there was no response or progression in the remainder. Complete tumor resection was achieved in 14 patients (36%). Twenty patients (51%) never became operable. Overall survival at 5 years was 28%, and event-free survival was 17%. Most deaths resulted from tumor progression (26 of 28). Presence of metastases and pretreatment extent of disease system grouping at diagnosis had an adverse inuence on overall survival in multivariate analysis. Conclusion: Survival for pediatric HCC patients is signicantly inferior to that for children with hepatoblastoma. Complete tumor excision remains the only realistic chance of cure, although it is often prevented by advanced disease. The presence of metastases is the most potent predictor of poor prognosis. A prospective worldwide cooperation in the eld of pediatric HCC should be encouraged to look for novel therapeutic concepts. J Clin Oncol 20:2798-2804. 2002 by American Society of Clinical Oncology.

EPATOCELLULAR CARCINOMA (HCC) in childhood is rare and accounts for less than 0.5% of all pediatric malignancies; however, it is the second most common malignant liver neoplasm (after hepatoblastoma [HB]) in children.1,2 Its relative frequency is 0.5 to 1.0 cases per million children.2,3 It is more frequently encountered in older children and teenagers than in infants. The most common clinical sign is hepatomegaly, which is sometimes associated with abdominal pain. Unlike HB, HCC often develops in the presence of underlying liver disease and cirrhosis, especially viral hepatitis (mainly hepatitis B) or inborn metabolic errors, such as tyrosinemia.1,4,5 Compared with HB, the prognosis in HCC is poor.1,6-9

From the Department of Pediatric Surgery, Medical University of Gdan sk, Gdan sk, Poland. Submitted June 18, 2001; accepted March 4, 2002. Supported by the Swiss Cancer League, Bernese Cancer League, Berno, Switzerland, and the United Kingdom Liver Tumor Parents Group. Address reprint requests to Piotr Czauderna, MD, Department of Pediatric Surgery, Medical University of Gdan sk, Ul, Nowe Ogrody 1-6, 80-803 Gdan sk, Poland; email: pczaud@amg.gda.pl. 2002 by American Society of Clinical Oncology. 0732-183X/02/2012-2798/$20.00

The Group for Epithelial Liver Tumors of the International Society of Pediatric Oncology (SIOPEL) was formed in 1989 and in 1990 began its rst clinical trial devoted to both pediatric HB and HCC. The main clinical aim of the trial was to improve survival and reduce operative morbidity and mortality in children with primary epithelial liver tumors by using preoperative chemotherapy with a combination of the most effective documented single agents (doxorubicin and cisplatin [PLADO]). The primary scientic objectives were to collect information on the epidemiology and natural history of these tumors and to identify prognostic factors. The results of the treatment of HB patients in the SIOPEL 1 trial were published in the Journal of Clinical Oncology in 2000.10 Several reports on the treatment of pediatric HCC have been published, but most of them either present a series less than 10 patients or combine patients treated in different ways1,4,6-9,11-13. The aim of this article is to present the results for patients with HCC treated in SIOPEL 1 and to compare them with the SIOPEL 1 HB patients and other published HCC trials. To the best of our knowledge, ours is the largest series of uniformly treated patients ever reported.
PATIENTS AND METHODS
The SIOPEL 1 trial was an international, prospective, cooperative single-arm study. It was open to patient registration between January

2798

Journal of Clinical Oncology, Vol 20, No 12 (June 15), 2002: pp 2798-2804 DOI: 10.1200/JCO.2002.06.102 Downloaded from jco.ascopubs.org on March 20, 2014. For personal use only. No other uses without permission. Copyright 2002 American Society of Clinical Oncology. All rights reserved.

HCC IN CHILDREN: SIOPEL 1 EXPERIENCE 1990 and February 1994 and involved 91 centers from 30 countries of the world. Children aged less than 16 years old with a histologically proven, previously untreated HB or HCC were eligible for the study. Central review of the pathology slides was required for each patient entering onto the study. Details of study design, data collection, and statistical methods and packages used have already been published together with the outcome for patients with HB.10,14

2799

Required Patient Information at Diagnosis and Staging Procedures

The determination of the serum alpha-fetoprotein (AFP) concentration value was mandatory at diagnosis. Open or closed surgical biopsy was strongly recommended for accurate diagnosis before chemotherapy, except in those patients for whom the surgical risk was considered unacceptable in the face of unequivocal clinical ndings (hepatic mass and increased AFP). Biopsy was mandatory in children younger than 6 months of age. All material obtained at biopsy and resection was to be centrally reviewed by one pathologist (J.K.).

Staging Procedures and Pretreatment Extent of Disease System (PRETEXT)

The pretreatment assessment of the extent of the primary tumor was by abdominal ultrasound, computed tomography (CT) scan, or magnetic resonance imaging. Metastatic spread was assessed by a chest x-ray (posteroanterior and lateral views) and, where available, lung CT scan. Hepatic angiography was optional. Imaging data sheets were reviewed centrally. Based on the radiologic ndings, a Pretreatment Extent of Disease System (PRETEXT) was developed to describe tumor extent at diagnosis, before therapy, and during therapy (Fig 1). Extension of the tumor beyond the liver was indicated with a V for extension into the vena cava and/or all three hepatic veins, P for extension in the main or both left and right branches of the portal vein, E for extrahepatic extension except for P and V, and M for the presence of distant metastases.

Treatment Plan and Chemotherapy Details

All patients were scheduled to receive the same treatment, consisting of preoperative PLADO chemotherapy, every 3 weeks and, where feasible, resection of the primary tumor after four, ve, or at most six PLADO courses. Each course of PLADO chemotherapy comprised cisplatin 80 mg/m2 on day 1 in a continuous 24-hour intravenous infusion and doxorubicin 30 mg/m2/d as a continuous 48-hour intravenous infusion on days 2 and 3 (total dose per course, 60 mg/m2). No more than six courses of PLADO were to be administered to each patient. The total planned dose of cisplatin was therefore 480 mg/m2 and for doxorubicin, 360 mg/m2. The aim of limiting the total dose of both drugs was to reduce the incidence of cardiotoxicity from doxorubicin and nephropathy and ototoxicity from cisplatin. Radiation therapy was permitted for small-volume abdominal disease that could not be safely resected.

Fig 1.

Schematic representation of the PRETEXT system.

Timing of Tumor Response Evaluation and of Surgery

Tumor response was evaluated after the second and fourth courses of PLADO by repeating the imaging investigations that were positive at diagnosis and monitoring the serum AFP level. If, as dened in the next section, there was evidence of progressive disease at either time point, the patient was switched to alternative chemotherapy. After four courses of PLADO, tumor resectability was assessed. If feasible, partial hepatectomy was then performed, and two more postoperative PLADO

courses were given. If the tumor was responding to chemotherapy but was still considered unresectable, two more courses of PLADO were given, and tumor resectability was reassessed.

Tumor Response
Complete response was dened as no evidence of disease and normal serum AFP. Partial response was dened as any tumor volume shrinkage (ie, decrease in at least one dimension with other dimensions decreased or stable) associated with a decreasing AFP value. Stable disease was dened as no change in tumor volume or serum AFP value.

Downloaded from jco.ascopubs.org on March 20, 2014. For personal use only. No other uses without permission. Copyright 2002 American Society of Clinical Oncology. All rights reserved.

2800
Progressive disease was an unequivocal increase in one or more dimensions or the appearance of new lesions.

CZAUDERNA ET AL

Surgery
No specic guidelines were provided for resection of the tumor, but to minimize the risk of perioperative deaths, it was recommended that: (1) the nutritional status of the patient be evaluated carefully before surgery, as described by Fan et al,15 and (2) patients be treated in centers with considerable experience in pediatric liver surgery and anesthesia. Primary surgery was admissible only in those patients in which the tumor was conned to the left lateral or right posterior sector (PRETEXT I). Orthotopic liver transplantation was considered in patients whose tumor was completely conned to the liver and that, despite a positive response to adequate rst line chemotherapy, remained unresectable. Results of orthotopic liver transplantation in the SIOPEL 1 study for HCC and HB will be presented in a separate manuscript. No specic guidelines were provided for patients with lung metastases, except that lung surgery was recommended, if feasible, in cases where pulmonary lesions persisted after PLADO therapy.

Statistical Methods
The results of this trial are expressed in terms of response to PLADO, overall survival, and event-free survival. All calculations are based on intention to treat and include data on patients who died before any treatment could be given or during the rst weeks of chemotherapy. To be assessable for response, patients must have received at least two courses of chemotherapy. Overall survival time was dened as the time interval between the date of diagnosis and the date of death (from any cause) or the date of last follow-up. Event-free survival was dened as the time interval from the date of diagnosis to one of the following: (1) date of alternative treatment because of failure of the PLADO regimen, (2) date of relapse, (3) date of death, or (4) date of last follow-up, whichever occurred rst. Date of alternative treatment was used as a surrogate for date of progression when the precise date of progression was not recorded. Relapse was identied from unequivocal imaging and serially increasing serum AFP levels or biopsy conrmation. Patients lost to follow-up were censored in these analyses. The Kaplan-Meier method was used to estimate survival curves.9 Ninetyve percent condence intervals (95% CIs) were calculated for the survival estimates using the Greenwood method.10 Variables at diagnosis were univariately assessed using the log-rank test for association with overall survival or event-free survival. Variables were considered to be statistically correlated at the 10% level of signicance. Variables correlated with survival were then considered in a multivariate analysis. The stepwise Cox proportional hazards regression model was used to model the prognostic factors adjusted for the inuence of each other. All statistical procedures were carried out using the SAS statistical package (SAS Institute, Cary, NC).

(31%) of 39 children, extrahepatic tumor extension was found in seven patients (18%), and major vascular involvement (all three hepatic veins, vena cava, and/or both branches of portal vein) was present in eight cases (21%). Most tumors were classied as the epithelial histologic variant of HCC (29 of 39; 75%), six of 39 tumors were brolamellar, and the rest (n 4) had other histology, ie, poorly differentiated or clear cell. Tumor staging revealed that a PRETEXT III and IV category were most common (11 and 13 cases, respectively; 61%), whereas PRETEXT I was found in one case and PRETEXT II in 14 patients (36%). Serum AFP concentration at diagnosis ranged from 1 to 1,400,000 ng/mL (median, 9,677 ng/mL). Serum AFP was normal ( 10 ng/mL) in 12 children (31%). All patients with a brolamellar HCC variant had low AFP levels ( 25 ng/mL). Tumor diameter varied from 7 to 26 cm (median, 14 cm). Multifocal tumors were present in 22 patients (56%), four of whom liver involvement was diffuse with no healthy hepatic parenchyma seen on the CT. Thirteen patients (33%) presented with HCC associated with hepatitis B and hepatic cirrhosis. One patient had tyrosinemia, and one patient had biliary cirrhosis. Tumor biopsy was performed at diagnosis in 34 of 39 cases (open in 12 and closed in 22). In two others, the diagnosis was made on clinical grounds, and in one case, data on biopsy are missing. In all three cases, the clinical diagnosis of HCC was conrmed by the resection specimen. In the remaining two patients, primary tumor resection was performed because of local center decision. Response to Preoperative Chemotherapy Thirty-seven patients received preoperative chemotherapy of one to six PLADO courses (median, four courses). Partial response was observed in 18 (49%) of 37 patients. No response was seen in eight (22%) of 37 patients, and disease progressed in six (16%) of 37 patients. For one patient, who was later lost to follow-up, data on response are missing. Four patients received only one PLADO course, therefore their response to chemotherapy could not be judged. Surgery Tumor resection was attempted in 19 patients (including 17 [46%] of 37 patients who received preoperative chemotherapy). It was complete in 14 of them, including two total hepatectomies with subsequent liver transplantations, although in one case, the transplant had been performed because of underlying tyrosinemia and not because the tumor itself was unresectable. Thus, the rate of successful tumor excision was 36%. Ten patients with completely

RESULTS

Patient and Tumor Characteristics Forty patients diagnosed with HCC were registered onto the SIOPEL 1 study, but one was excluded from further analysis because of insufcient data. Male to female ratio was 2.5:1. Age range was from 4 to 15 years (median, 12 years). At diagnosis, lung metastases were identied in 12

Downloaded from jco.ascopubs.org on March 20, 2014. For personal use only. No other uses without permission. Copyright 2002 American Society of Clinical Oncology. All rights reserved.

HCC IN CHILDREN: SIOPEL 1 EXPERIENCE

2801

Fig 2.

Overall survival of SIOPEL 1 HCC patients.

Fig 3.

Event-free-survival of SIOPEL 1 HCC patients.

resected tumors responded to preoperative chemotherapy, and two had primary tumor surgery. The 13th patient had stable disease after preoperative chemotherapy, whereas data on chemotherapy response of the 14th patient are missing. Partial excision of the tumor (with microscopic residual) was performed in two patients, one of whom had local radiotherapy to residual disease, and in three patients, the tumor was inoperable at laparotomy. The following types of liver resection were performed: (1) eight hemihepatectomies, (2) ve extended hemihepatectomies, (3) two total hepatectomies, and (4) one left lateral lobectomy. In 20 cases of HCC (51%), primary tumors never became operable because of extensive liver involvement. Survival Eight patients (28%) are alive with no evidence of disease at the median follow-up time of 75 months (from 49 to 90 months). Among them there is one child with brolamellar HCC variant. All survivors underwent complete tumor resection. Twenty-seven children (69%) died of progressive disease. One child with progressive disease died of chemotherapy complications after secondline chemotherapy with carboplatin and etoposide. Another child died after complete tumor resection because of surgical complications (septic shock). Two patients were lost to follow-up. The median survival of the patients who died was 15 months (range, 1 to 65 months). Six relapses occurred (all after complete resections). Three relapses were local, two were distant, and one was combined. Four relapsed patients died of progressive disease, one patient is alive in second remission, and one was lost to follow-up. The gures for overall survival for HCC patients in the SIOPEL 1 study are 40% (95% CI, 24% to 55%) at 2 years and 28% (95% CI, 14% to 43%) at 5 years (Fig 2). Event-free survival at 2 years was 23% (95% CI, 10% to

37%), whereas it was 17% (95% CI, 6% to 30%) at 5 years (Fig 3). Factors Inuencing Prognosis Five factors at diagnosis were considered for their relationship with overall survival and event-free survival: metastases, focality of tumor, enlargement of the hilar nodes, vascular invasion (identied from CT or ultrasound), and PRETEXT grouping. Three factors were statistically signicant for overall survival at the 10% level and they are metastases, PRETEXT grouping, and focality of tumor. The results of this analysis are listed in Table 1. The results are also shown by Kaplan-Meier curves (Figs 4, 5, and 6). In the multivariate analysis, all three of the above factors were entered into the model, and PRETEXT and metastases were identied as predictors of overall survival (hazard ratios, 0.16 [95% CI, 0.04 to 0.68] and 1.82 [95% CI, 1.01 to 3.2], respectively). Initial PRETEXT discriminated survival groups so well that it could provide valuable patient stratication at diagnosis (Fig 6).

Table 1.

Factors Signicant for Overall Survival

5-Year Overall Survival

Parameter

df

95% CI

PRETEXT I-II III IV Metastases Yes No Multifocality Unifocality

6.6

.09

44 22 8 9 37 11 45

16-72 0-49 0-28 0-26 19-57 0-26 21-70

15 4.52

1 1

.0001 .03

Downloaded from jco.ascopubs.org on March 20, 2014. For personal use only. No other uses without permission. Copyright 2002 American Society of Clinical Oncology. All rights reserved.

2802

CZAUDERNA ET AL

Fig 4. Overall survival of SIOPEL 1 HCC patients and presence of metastases.

Fig 6.

Overall survival of SIOPEL 1 HCC patients and PRETEXT category.

For event-free survival, metastases and vascular invasion were signicant at the 10% level in the univariate analysis (P .0001 and P .08, respectively; Table 2). In the multivariate analysis, the only predictor of event-free survival was metastases (hazard ratio, 0.167 [95% CI, 0.04 to 0.6]).
DISCUSSION

This is the rst study reporting such a large series of uniformly treated (in a prospectively planned fashion) pediatric HCC patients. In all HCC reported series, therapeutic results are equally poor with overall survival less than 30%.1,4,6-9,11,16 Also, in this study, treatment results were unsatisfactory (28% overall survival and 17% eventfree survival), although they have been among the best ever reported.1,4,9,10,12,16 In our series, 33% had underlying liver disease. This observation conrms that the majority of pediatric hepatocarcinomas are de novo cases, usually not related to hepatic cirrhosis, in contrast to the adult experience.1,5,17-19 We could not conrm earlier adult observations on better outcome associated with brolamellar variant of HCC, although there were only a few such patients in the reported

group (six of 39; one is alive with no evidence of disease, one was lost to follow-up, and four of them died).1,20,21 However, median survival of patients with brolamellar HCC was much longer (25 months v 11 months) than that for the rest of HCC group. Optimal therapy for pediatric HCC patients is yet to be determined. The advanced stage of most tumors at diagnosis (as far as local PRETEXT grouping, presence of metastases, and multifocality are concerned) and the relative insensitivity of HCC to current chemotherapy mean that the rate of complete excision is low (36% in the reported study compared with 87% for HB in the same SIOPEL 1 study), which is in line with other reports.1,4 All authors agree that complete tumor excision remains a cornerstone of therapy (in our series, eight of 14 patients who had complete resection are alive, and they are the only survivors).1,4,16 In our series, multifocal tumors strongly inuenced the possibility of surgical resection (10% complete resections in multifocal tumors v 61% in unifocal tumors). This is probably why multifocality of the tumor was one of the identied prognostic factors. Presence of vascular invasion on imaging was another one, and a similar observation was made by Tagge et al.16 Both multifocality and vascular invasion could be an explanation for the much higher relapse rate and lower survival in completely resected HCC patients in
Table 2. Factors Signicant for Event-Free Survival
5-Year Event-Free Survival Parameter

df

95% CI

Fig 5. Overall survival of SIOPEL 1 HCC patients and tumors multifocality.

Metastases Yes No Vascular invasion Yes No

17.4

.0001

0 25 0 27

0-0 8-41 0-0 4-49

3.06

.08

Downloaded from jco.ascopubs.org on March 20, 2014. For personal use only. No other uses without permission. Copyright 2002 American Society of Clinical Oncology. All rights reserved.

HCC IN CHILDREN: SIOPEL 1 EXPERIENCE

2803
many adult studies (usually 50% decrease in the sum of the products of the tumors longest dimension).24-26 None of the patients who did not respond to preoperative chemotherapy survived. Thus, it seems that chemotherapy response in HCC may be used as an additional indicator of the chance of cure. The relative chemoresistance of HCC is also reected by the fact that metastatic patients had the worst prognosis, and the presence of metastases was the most potent adverse prognostic factor. This nding was conrmed by Chen et al.4 It is speculated that refractoriness of HCC to chemotherapy results from the presence of multidrug resistance genes, including p53 gene.27 Overall survival in pediatric HCC is far from satisfactory, and the relative resistance of this neoplasm to present therapeutic regimens indicates that new treatment approaches are required. It could well be that the whole strategy of treatment should be redesigned. Some new chemotherapeutic agents (for example, paclitaxel or antiAFP monoclonal antibodies) have already been tested in children and adults but with contradictory results.28-30 Other new treatment strategies that are being investigated are transarterial chemoembolization, cryosurgery, thermotherapy, or large-dose intra-arterial chemotherapy under conditions of hepatic venous isolation and hemoperfusion.2426,30-33 However, many of these treatments, at least in adults, are only palliative or at best prolong life rather than achieve cure. Others, like chemoembolization, are more promising but difcult to test in the framework of multicenter, international trials. Survival for pediatric HCC patients is signicantly inferior to that for children with HB despite similar multidisciplinary treatment. At present, complete tumor excision remains the only realistic chance of cure. However, in many cases it cannot be achieved because of advanced disease at diagnosis and refractoriness of HCC to standard chemotherapy. Because pediatric HCC is relatively rare in the Western world, most groups incorporate these patients into HB trials. In light of the above observations, this approach should now be questioned. It is probably more appropriate to encourage a prospective worldwide cooperation in the eld of pediatric HCC to look for and test new treatment approaches and novel therapeutic concepts.
ACKNOWLEDGMENT
We thank Claire Dicks-Mireaux, MD, for her invaluable work with imaging in SIOPEL 1 study.

comparison with HB patients (5-year disease-free survival in SIOPEL 1 was 92% for completely resected HB v 54% for HCC). It is also possible that multifocality is even more frequent, but small lesions are not picked up on imaging and remain undetected and hence unresected. If this hypothesis is true, these unidentied lesions may be to blame for the higher proportion of relapses in HCC. It is difcult to analyze separately the patients who relapsed because they are a small subset of children. Their median age, tumor size, and frequency of liver cirrhosis were similar to the whole group. The majority of them were unifocal (ve of six) and in the PRETEXT II group (ve of six), but that was probably because most multifocal and more advanced tumors were unresectable (90%). It is worth noting that ve (83%) of six relapsed HCC patients had low AFP levels ( 100 ng/mL) compared with three (33%) of nine who did not relapse. Half of the relapsed children (three of six) apart from the low AFP levels had brolamellar HCC variant. Other reports indicate even lower survival rates in completely excised cases of approximately 30% to 40%.1,4,6 The role of liver transplantation in pediatric HCC is still very much a matter for debate; however, in adults, transplantation criteria for HCC are strict, including disease conned to the liver with no more than three tumor foci of a maximum diameter of 3 cm.16,22 In general, orthotopic liver transplantation is not an option for most HCC patients because of advanced disease at diagnosis and the much higher proportion of recurrences in comparison with HB.16 This issue will be discussed in detail in a separate article devoted to transplantation in SIOPEL 1. Chemotherapy is only partially effective for HCC (approximately 50% partial responses in the reported series). Although the response to PLADO chemotherapy in this study constitutes a signicant improvement in response rate in comparison with historical controls, our experience has been that observed responses are not sufcient to render the majority of patients resectable, which is in parallel with other reports.4,23 Nevertheless, resectability rate and survival were much higher among patients who responded to preoperative chemotherapy (61% v 5% and 33% v 0%, respectively). The response to chemotherapy achieved in this trial seems to be a signicant improvement in comparison with other reports.4,16 However, it has to be mentioned that the response denition used in this protocol was different (ie, any tumor shrinkage and decrease of AFP) to that used by

Downloaded from jco.ascopubs.org on March 20, 2014. For personal use only. No other uses without permission. Copyright 2002 American Society of Clinical Oncology. All rights reserved.

2804
APPENDIX

CZAUDERNA ET AL

The appendix of study participants is available online at www.jco.org.

REFERENCES
1. Moore SW, Hesseling PB, Wessels G, et al: Hepatocellular carcinoma in children. Pediatr Surg Int 12:266-270, 1997 2. Bellani FF, Massimino M: Liver tumors in childhood: Epidemiology and clinics. J Surg Oncol Suppl 3:119-121, 1993 3. Parkin DM, Stiller CA, Draper GJ, et al: The international incidence of childhood cancer. Int J Cancer 42:511-520, 1988 4. Chen JC, Chen CC, Chen WJ, et al: Hepatocellular carcinoma in children: Clinical review and comparison with adult cases. J Pediatr Surg 33:1350-1354, 1998 5. Di Bisceglie AM, Rustgi VK, Hoofnagle JH, et al: NIH conference: Hepatocellular carcinoma. Ann Intern Med 108:390-401, 1988 6. von Schweinitz D, Fuchs J, Mildenberger H: Surgical strategy in pediatric liver malignancies. Langenbecks Arch Chir Suppl Kongressbd 113:1091-1094, 1996 7. Exelby PR, Filler RM, Grosfeld JL: Liver tumors in children in the particular reference to hepatoblastoma and hepatocellular carcinoma: American Academy of Pediatrics Surgical Section Survey 1974. J Pediatr Surg 10:329-337, 1975 8. von Schweinitz D, Burger D, Weinel P, et al: Therapy of malignant liver tumors in childhood: An intermediate report of the HB-89 multicenter study of the GPOH. Klin Padiatr 204:214-220, 1992 9. Giacomantonio M, Ein SH, Mancer K, et al: Thirty years of experience with pediatric primary malignant liver tumors. J Pediatr Surg 19:523-526, 1984 10. Pritchard J, Brown J, Shafford E, et al: Cisplatin, doxorubicin and delayed surgery for childhood hepatoblastoma: A successful approachResults of the rst prospective study of the International Society of Pediatric Oncology. J Clin Oncol 18:3819-3828, 2000 11. Gauthier F, Valayer J, Thai BL, et al: Hepatoblastoma and hepatocarcinoma in children: Analysis of series of 29 cases. J Pediatr Surg 21:424-429, 1986 12. Schmidt A, Schmidt P, Haas RJ, et al: Primary liver tumors in childhood: An analysis of 19 cases. Langenbecks Arch Chir 373:159165, 1988 13. Gururangan S, OMeara A, MacMahon C, et al: Primary hepatic tumours in children: A 26-year review. J Surg Oncol 50:30-36, 1992 14. Brown J, Perilongo G, Shafford E, et al: Pretreatment prognostic factors for children with hepatoblastoma: Results from the International Society of Paediatric Oncology (SIOP) studySIOPEL 1. Eur J Cancer 36:1418-1425, 2000 15. Fan ST, Lo CM, Lai EC, et al: Perioperative nutritional support in patients undergoing hepatectomy for hepatocellular carcinoma. N Engl J Med 331:1547-1552, 1994 16. Tagge EP, Tagge DU, Reyes J, et al: Resection, including transplantation, for hepatoblastoma and hepatocellular carcinoma: Impact on survival. J Pediatr Surg 27:292-297, 1992 17. Weinberg AG, Finegold MJ: Primary hepatic tumours of childhood. Hum Pathol 14:512-537, 1983 18. Kew MC: Hepatocellular carcinoma with and without cirrhosis. Gastroenterology 97:136-139, 1989 19. Ducreux M, Elias D, Rougier P, et al: Treatment of hepatocellular carcinoma in the presence of liver cirrhosis. J Chir (Paris) 132:279-286, 1995 20. Farhi DC, Shikes RH, Murari PJ, et al: Hepatocellular carcinoma in young people. Cancer 52:1516-1525, 1983 21. Saab S, Yao F: Fibrolamellar hepatocellular carcinoma: Case reports and a review of the literature. Dig Dis Sci 41:1981-1985, 1996 22. Superina R, Bilik R: Results of liver transplantation in children with unresectable liver tumors. J Pediatr Surg 31:835-839, 1996 23. Douglass EC: Hepatic malignancies in childhood and adolescence (hepatoblastoma, hepatocellular carcinoma and embryonal sarcoma). Cancer Treat Res 92:201-212, 1997 24. Ku Y, Fukumoto T, Iwasaki T, et al: Clinical pilot study on high-dose intraarterial chemotherapy with direct hemoperfusion under hepatic venous isolation in patients with advanced hepatocellular carcinoma. Surgery 117:510-519, 1995 25. Ravikumar TS, Pizzorno G, Bodden W, et al: Percutaneous hepatic vein isolation and high-dose hepatic arterial infusion chemotherapy for unresectable liver tumors. J Clin Oncol 12:27232736, 1994 26. Curley SA, Newman RA, Dougherty TB, et al: Complete hepatic venous isolation and extracorporeal chemoltration as treatment for human hepatocellular carcinoma: A phase I study. Ann Surg Oncol 1:389-399, 1994 27. Shen DW, Lu YG, Chin KV, et al: Human hepatocellular carcinoma cell lines exhibit multidrug resistance unrelated to MRD1 gene expression. J Cell Sci 98:317-322, 1991 28. Chao Y, Chan WK, Birkhofer MJ, et al: Phase II and pharmacokinetic study of paclitaxel therapy for unresectable hepatocellular carcinoma patients. Br J Cancer 78:34-39, 1998 29. Hata Y, Takada N, Sasaki F, et al: Immunotargeting chemotherapy for AFP-producing pediatric liver cancer using the conjugates of anti-AFP antibody and anti-tumor agents. J Pediatr Surg 27:724-727, 1992 30. Whang-Peng J, Chao Y: Clinical trials of HCC in Taiwan. Hepatogastroenterology 45:1937-1943, 1998 31. Trinchet JC, Beaugrand M: Hepatocellular carcinoma: Treatment with arterial chemo-embolization. Presse Med 14:831-833, 1994 32. Adam R, Akpinar E, Johann M, et al: Place of cryosurgery in the treatment of malignant liver tumors. Ann Surg 225:39-48, 1997 33. Venook AP: Treatment of hepatocellular carcinoma: Too many options? J Clin Oncol 12:1323-1334, 1994

Downloaded from jco.ascopubs.org on March 20, 2014. For personal use only. No other uses without permission. Copyright 2002 American Society of Clinical Oncology. All rights reserved.